Skip to content

Dataset · mteb

Compact search embedding evaluation benchmark

mteb/NanoMSMARCORetrieval

A lightweight subset of the famous MS MARCO search dataset for fast embedding model evaluation in MTEB.

What's inside

The dataset contains from 1 to 10 thousand English text records taken from the MS MARCO web corpus. It is part of the popular Massive Text Embedding Benchmark (MTEB) and is designed for fast evaluation of text-retrieval quality.

What you can do with it

  • Fast evaluation and comparison of text embedding models in the MTEB library
  • Debugging and validation of semantic search pipelines and RAG systems
Tasks
Text retrieval
Languages
English
Size
1K–10K rows
Downloads this month
22.4K
Likes
0

How to load it

from datasets import load_dataset

ds = load_dataset("mteb/NanoMSMARCORetrieval")